According to a report published by the Wall Street Journal this Wednesday (14), Microsoft and OpenAI experienced some awkwardness before the early launch of the Bing chatbot in February.
When it was released, the Bing was vulnerable to prompt injection attacksrevealing company secrets and providing sometimes inaccurate and truly unbalanced responses to user prompts.
According to the WSJ, OpenAI warned Microsoft “about the dangers of rushing to integrate OpenAI technology without further training” and “suggested Microsoft to be slower in integrating its AI technology with Bing”. One of OpenAI’s main concerns was that Bing’s chatbot Sydney could give inaccurate or unbalanced responses, but this initial warning was ignored by Microsoft.
According to Microsoft CEO Satya Nadella, the flaws were part of Microsoft’s plan to train the chatbot to respond to real-world requests that could not be tested in the lab.
As OpenAI started public testing of ChatGPT while Microsoft was still working on integrating OpenAI technology into Bing, tension apparently rose between the partners, who were also rivals in an artificial intelligence race. Analytics firm YipitData reported that ChatGPT reached 200 million monthly users, while Bing reached 100 million daily active users in March.
According to the WSJ, more tension and confusion arose within Microsoft’s internal AI team, who “complained about decreasing spending”. Most employees are hampered by the lack of “access to the inner workings” of the OpenAI technology, which complicates the routine of those who want to integrate this technology into various Microsoft products.
Much of this conflict amounts to typical internal disputes that happen whenever two companies come togetherreported the WSJ, but there’s no ignoring the conflict inherent in both sides trying to maintain independence while making maximum profits by selling access to the same technology.
Despite these tensions, Nadella told Wired that OpenAI “went for” Microsoft and Microsoft “went for” OpenAI. He still envisions “a good commercial partnership” between the independent companies and considered Microsoft’s investment in OpenAI as “a stable long-term agreement”.
